Building a Sound Credit Ecosystem:  Benefits of data-driven decision-making


Building a Sound Credit Ecosystem

Benefits of data-driven decision-making


In the film Interstellar, the hero looked down at the world’s past and present when traveling through multi-dimensional space and finally understood the meaning of life. The same applies to the credit eco-system. "A healthy credit ecosystem can neither be established overnight nor with individual effort, and it can never be considered a system established once and for all. It requires each role in the system to be forward-looking and aware of the big picture," said Jerry Yu, Partner, Enterprise Risk Services, Deloitte.

Currently, the credit ecosystem is growing rapidly and has fierce competition. If individual users are one-dimensional dots, and business transactions between users and counterparties are two-dimensional lines, the credit industrial chain is a three-dimensional plane, and the credit ecosystem a four-dimensional object. The key to cutting through confusion is embedding an integrated, symbiotic, and changing development strategy into the whole ecosystem. Backed by advanced credit risk management techniques, we should be forward-looking and survey this four-dimensional world from a macro-perspective; we should establish a new, optimized system so that a harmonious and symbiotic environment can emerge.

Industry demand

It is the best of times; it is also the worst of times. Like oil during the Second Industrial Revolution, big data in the Internet Age has been underutilized. From 2009 to 2020, data of financial markets are expected to grow by 4300 percent to 35 Zettabytes, yet only 34 percent of data have been utilized by financial institutions. Big data, with prediction at its core, is valuable in precision marketing, operation optimization, client management, real-time processing, and risk assessment. It transforms decision making from being business-driven to being data-driven in all industries.

With the development of emerging businesses in financial markets, there will be an increase in clients' transactions as well as demand for payment, investment, and financing. At the same time, transaction products and transaction structure are increasingly complex. Information obtained in a traditional way from financial institutions and credit agencies can no longer meet the market's requirements. Besides, with the opening of financial markets, the government will no longer endorse the bail-out trust, which should lead to a surge in default both in terms of numbers and amounts. In such a harsh market environment, there will be more of an emphasis on the accuracy and promptness of ratings. These changes are putting new demand on rating methods and data. In addition to the traditional financial data, new data will gradually be included in credit risk assessment, such as text messages and conversations via mobile phones, shopping information, and social platform information. Compared with traditional ratings, which entirely rely on enterprises' financial data, market positions, and governance transparency, the new rating idea is revolutionary. By analyzing these valuable data sources, rating service providers can fully explore client information and transform it into credit scores to have an accurate understanding of the market and fully explore clients' value.

One-dimensional credit, two-dimensional transactions, and a three-dimensional industry chain

Credit is a special form of value movement based on loan repayment within agreed terms and interest payments. The essential credit rating is comprised of three parts.

  • Users: Have investment and financing demands, including service users such as individual and corporate users, and service providers such as financial institutions. Rate the principal and debts with users’ service intention.
  • Data providers: Provide user demand data for rating agencies.
  • Rating service providers: Provide services in credit management products and consulting.
Click to enlarge image

The closed-loop credit rating industry chain has a mutual restriction and interdependence relationship. Rating agencies create credit risk rating frameworks and evaluate client principals and projects using financial and non-financial data. Their precise evaluations rely on years of data accumulation, understanding of financial markets, analysis of rating targets, and an accurate grasp of the macro economy. An excellent rating system better enables financial institutions to identify clients’ quality and develop tailored strategies in order to maximize the value of clients’ businesses and optimize the services of financial institutions.

Four-dimensional credit ecosystem

A credit ecosystem is a place for all players to coexist in the credit chain. Individuals in the ecosystem do not compete with but instead mutually depend on and reinforce each other.

Click to enlarge image

Three levels of interrelationship among individuals in the credit ecosystem:

  • First level: the business relationship among different roles

Taking banks as an example, as the industry chain shows, service providers need to evaluate counterparties’ repayment capabilities. Without evaluation capability, banks would have to consult with external rating agencies. Moreover, banks themselves have access to a sea of data, including customers’ basic data, transaction data, and clearing and settlement data, which can be used as core data in evaluation, and can also be given to rating agencies for credit modeling after data masking. Without enough data, banks would need to purchase data from external data providers, such as Bloomberg, Reuters, Wind, and other third-party Internet platforms. Having been entrusted by banks, rating agencies would build a credit risk model to import client data and perform quantitative analysis with these data, finally producing credit scores.

  • Second level: the cooperation relationship among similar roles

Second level relationships mainly lie in information exchanges among users, resource integration among data providers, and technology sharing of rating service providers.

The integration of information and data resources should benefit the sustainable development of the credit ecosystem, especially the resource concentration of Big Data. Big Data is composed of structured data and non-structured data. Structured data is mainly controlled by data providers and SOEs, such as customer transaction data, juridical data, and water, power, coal, gas, and tax data. Non-structured data is mainly Internet behavior data. Only by integrating data from different sources can we fully exploit the value of data and offer superior services. For instance, in 2015, eight private agencies were the first batch to be granted customer credit licenses. These agencies, who have collected large amounts of credit records, are also data providers. By integrating the data of traditional national agencies and financial institutions, we can build a national credit system and a credit data market in a better overall manner.

Technology sharing among rating service providers will also contribute to the healthy development of the credit ecosystem. Although model building and indicator selecting are both based on mathematic theories, when building the models we still need to consider model users’ risk preferences and business practices. A good rating model should fully consider the professional advice of relevant business departments on the basis of data analysis, so as to select the best model that can reflect the risk status of enterprises. Only by taking different perspectives into account to perfect the model can we meet rapidly developing business demands.

  • Third level: the overall coexistence relationship

The information and technology revolution brought by the Internet, Big Data, and machine learning is changing our lifestyles. Traditional financial service models can no longer meet the dramatically changing needs of the capital market. The emergence of Lending Club in the United States has redistributed retail investors’ assets. Like Facebook, Airbnb, and Uber, it is characterized by a sharing economy, and is pioneering Internet finance. Under these circumstances, credit ratings should advance with the times, and no longer only serve banking, securities, and other traditional institutions. Ratings should instead apply to any place where transactions happen, at any time, to form an inclusive and integrated credit ecosystem. For example, currently in China's first-tier cities, it is impossible for those without urban registration certificates to overdraft in order to pay utility bills such as water, electric, and gas. However, if service providers could enter into the credit ecosystem, services would be more efficient and convenient.

In short, users' service needs will encourage rating service providers to develop more accurate risk models. Meanwhile, users are expected to be increasingly satisfied with the quality rating services, which is likely to bring about new businesses and more potential users for data providers to collect more data for exploration. Such a cycle should help attain the constant upgrading updating of the credit system and full sharing of credit ratings, which in turn should generate harmonious coexistence and win-win development for the credit ecosystem.


"Three factors comprise a basic credit system: data as the base, credit risk models as the tool, and credit rating application as the promotion." said Jerry Yu, Partner, Enterprise Risk Services, Deloitte

Such a three-layer pyramid structure is in compliance with the development strategy featuring fusion at the bottom, co-existence in the middle, and reform at the top. The credit ecosystem is dynamic and it requires multi-party cooperation, resource integration, information sharing, model innovation, and upgrading and updating of rating application to meet the demands of the rapidly-developing industry, and thereby achieve co-existence and win-win development among all parties in the industry. The work that Deloitte Enterprise Risk Services is doing for our financial services clients is looking forward to such a situation.

Did you find this useful?